NVIDIA DRIVE Powers Mercedes-Benz S-Class Autonomous Driving Platform

On January 29, NVIDIA announced that the new Mercedes-Benz S-Class equipped with MB.OS is built on the NVIDIA DRIVE Hyperion architecture and NVIDIA DRIVE AV Level 4 software, marking a significant step in NVIDIA DRIVE autonomous driving for luxury vehicles. The platform is designed to support future robotaxi services while maintaining a strong focus on safety, reliability, and scalability.

NVIDIA DRIVE AV software and real-world reliability

NVIDIA DRIVE AV delivers a full-stack automated driving system capable of managing rare, complex, and unpredictable driving scenarios while prioritizing safety. Instead of relying only on predefined rules, the system continuously evaluates surrounding conditions and determines the safest possible response. The software is trained using NVIDIA DGX computing systems and validated through large-scale simulations powered by NVIDIA Omniverse and Cosmos models. This combination allows extensive testing across edge cases before deployment, helping ensure dependable performance in real-world environments.

AI-driven perception, planning, and decision-making

Built on NVIDIA’s advanced AI technologies, NVIDIA DRIVE autonomous driving integrates perception, planning, and decision-making into a unified system tailored specifically for Mercedes-Benz vehicles and their unique sensor configurations. The platform enables vehicles to assess situations, compare multiple driving strategies, and select the safest course of action in real time.

Human-like driving intelligence

At the core of the system are NVIDIA Alpamayo open models, which allow the vehicle to reason through complex traffic situations step by step. This approach supports smoother, more natural driving behavior that closely mirrors human judgment while maintaining strict safety boundaries.

NVIDIA DRIVE Hyperion sensor architecture and redundancy

The Mercedes-Benz S-Class uses NVIDIA DRIVE Hyperion, which combines cameras, radar, and lidar sensors with redundant hardware to enhance system reliability. This multimodal sensor setup ensures continuous operation even if one component encounters an issue. The architecture follows a defense-in-depth safety strategy, incorporating redundant computing systems, diverse sensor inputs, and dual software stacks. These include AI-based autonomous driving software alongside a classical safety stack designed to keep the vehicle within defined safe limits under all conditions.

Safety framework supporting Level 4 autonomy

Developed under the NVIDIA Halos safety framework, the platform is engineered to eliminate single points of failure and maintain operational safety at every layer of the system. This framework is a key enabler for Level 4 autonomous driving, supporting high levels of automation while meeting strict safety and reliability requirements. By integrating NVIDIA DRIVE autonomous driving technology with MB.OS, the Mercedes-Benz S-Class sets a foundation for future autonomous mobility solutions, combining advanced AI, simulation-based validation, and a safety-first design approach tailored for next-generation luxury and robotaxi applications.